Reduction Kinetics of Graphene Oxide Determined by Electrical Transport Measurements and Temperature Programmed Desorption

J. Phys. Chem. C XXXX, xxx, 000 A

Reduction Kinetics of Graphene Oxide Determined by Electrical Transport Measurements and Temperature Programmed Desorption

Inhwa Jung,† Daniel A. Field,‡ Nicholas J. Clark,‡ Yanwu Zhu,† Dongxing Yang,† Richard D. Piner,† Sasha Stankovich,§ Dmitriy A. Dikin,§ Heike Geisler,| Carl A. Ventrice, Jr.,*,‡ and Rodney S. Ruoff*,† Department of Mechanical Engineering and the Texas Materials Institute, UniVersity of Texas at Austin, Austin, Texas, 78712, Department of Physics, Texas State UniVersity, San Marcos, Texas, 78666, Department of Mechanical Engineering, Northwestern UniVersity, EVanston, Illinois, 60208, and Insitute for EnVironmental and Industrial Science, Texas State UniVersity, San Marcos, Texas, 78666 ReceiVed: May 11, 2009; ReVised Manuscript ReceiVed: August 31, 2009

The thermal stability and reduction kinetics of graphene oxide were studied by measuring the electrical resistivity of single-layer graphene films at various stages of reduction in high vacuum and by performing temperature programmed desorption (TPD) measurements of multilayer films in ultrahigh vacuum. The graphene oxide was exfoliated from the graphite oxide source material by slow-stirring in aqueous solution, which produces single-layer platelets with an average lateral size of ∼10 µm. From the TPD measurements, it was determined that the primary desorption products of the graphene oxide films for temperatures up to 300 °C are H2O, CO2, and CO, with only trace amounts of O2 being detected. Resistivity measurements on individual single-layer graphene oxide platelets resulted in an activation energy of 37 ( 1 kcal/mol. The TPD measurements of multilayer films of graphene oxide platelets give an activation energy of 32 ( 4 kcal/mol.

1. Introduction of the oxide are still debated. Some of the functional groups proposed to exist at the surfaces of graphene oxide are epoxide Graphene oxide, which is an electrical insulator, shows (C-O-C), single-bonded on-top oxygen (C-O), and hydroxyl promise for use in several technological applications. For (C-OH) groups. In addition, carbonyl (CdO) groups are instance, individual, monolayer, graphene oxide platelets could expected to form at missing carbon atom defect sites of the be used as dielectric layers in nanoscale electronic devices. Since surface, and carboxylic (OdC-OH) groups, carbonyl groups, the electrical, optical, and mechanical properties of graphene and hydroxyl groups are expected to form on the edges of the oxide can be controlled by chemical modification, films platelets. Recent solid-state NMR studies of 13C-labeled graphite composed of layers of graphene oxide platelets may be used as 23 - oxide have shed further light on this topic. the active region of chemical sensors.1 4 In principle, graphene oxide films could also be used as a precursor for the formation Before graphene oxide can be used in most technological of large-scale graphene films by either thermal or chemical applications, it is important to understand its thermal stability reduction of the graphene oxide.5-11 and reduction kinetics. For instance, the decomposition tem- Graphene oxide is formed by extensive chemical oxidation perature of graphene oxide will govern which processing Publication Date (Web): October 2, 2009 | doi: 10.1021/jp904396j of graphite to form graphite oxide12-14 followed by exfoliation techniques can be used to manufacture nanoscale electronic to monolayer thick platelets by techniques such as sonication6,15,16 devices and sensors based on this material. Despite the potential 11 technological importance of this material, there have been Downloaded by UNIV OF TEXAS AUSTIN on October 2, 2009 | http://pubs.acs.org or slow-stirring in aqueous solution. It is a nonstoichiometric relatively few studies of the thermal stability of either graphite compound; therefore, the carbon-to-oxygen (C/O) ratio depends 5,24,25 7,11,26,27 on the method of preparation and by how much the material is oxide or graphene oxide. An early study of the reduced. Previous measurements of the C/O ratio of graphene thermal decomposition of graphite oxide using gas chromatog- oxide films by X-ray photoelectron spectroscopy (XPS) have raphy found that the primary desorption products are H2O, CO2, determined that the ratio of fully oxidized graphene oxide is and CO and that the decomposition occurs over a temperature - ° 24 2:1.17 The geometric structure of fully oxidized graphene oxide range of 80 180 C. These results indicate that graphene oxide is still an open question. Several different models have been is a rather fragile material and that since there is carbon loss proposed, with some of these models predicting a planar upon reduction, the structural integrity of the carbon backbone structure,18,19 while others favor a buckled structure.17,20-22 of the graphene will most likely be compromised upon Although there is general agreement that the chemical composi- decomposition. In fact, evidence for the generation of holes in thermally expanded graphite oxide has been discussed previ- tion of graphene oxide is primarily carbon, oxygen, and 28 hydrogen, the assignment and positions of the functional groups ously. In this study, the reaction kinetics and reduction process of * To whom correspondence should be addressed. E-mail: cventrice@ graphene oxide films have been investigated using two tech- txstate.edu (C.A.V.); [email protected] (R.S.R.). niques: monitoring the resistivity of a single-layer platelet of † University of Texas at Austin. graphene oxide at various stages of reduction in high vacuum, ‡ Department of Physics, Texas State University. § Northwestern University. and temperature programmed desorption (TPD) measurements | Institute for Environmental and Industrial Science, Texas State University. of multilayer films under ultrahigh vacuum (UHV) conditions. 10.1021/jp904396j CCC: $40.75  XXXX American Chemical Society B J. Phys. Chem. C, Vol. xxx, No. xx, XXXX Jung et al.

Figure 2. (a) SEM image of submonolayer coverage of slow-stirred graphene oxide platelets deposited on a Si(100) substrate and (b) optical image of a ∼9 ML thick ﬁlm composed of single-layer graphene oxide platelets deposited on 300 nm SiO2/Si(100).

Figure 1. Schematic of the experimental setup for (a) the resistivity measurements of an individual single-layer graphene oxide platelet and (b) the TPD measurements of multilayer graphene oxide ﬁlms.

A schematic of these two techniques is shown in Figure 1. For the first method, the resistivity is monitored at constant temperature as a function of time and is then repeated for different temperatures ranging from 140 to 200 °C. From the time dependence of the resistivity, the reaction order, rate Figure 3. (a) Optical image of patterned electrodes on an individual constant, and an estimated value of activation energy are graphene oxide platelet and (b) sample mounting arrangement for TPD acquired. From the TPD measurements, the relative coverage measurements of multilayer graphene oxide films. of the desorption products is determined by analyzing the areas under the partial pressure versus time curves. In addition, activation energies for decomposition of the oxide are deter- above, which contains large platelets of single-layer graphene mined with TPD by measuring the shift of the temperature at oxide, is used. This technique results in graphene oxide platelets which the maximum of the partial pressure versus temperature with an average lateral size of ∼10 µm. A scanning electron plot occurs (TM) as the heating rate () of the sample is microscope (SEM) image of a submonolayer coverage of slow- increased. From the TPD measurements, it is determined that stirred platelets on a Si(100) substrate is shown in Figure 2a. the primary desorption products of the graphene oxide films For the resistivity measurements, single-layer graphene oxide for temperatures up to 300 °C are H2O, CO2, and CO, with a films were formed by applying a droplet of the slow-stirred similar TM for all three molecular species. Activation energies graphene oxide solution on the substrate followed by dispersion of 37 ( 1 and 32 ( 4 kcal/mol are measured by resistivity of the droplet by blowing with dry N2 gas. The electrical measurements on single-layer graphene oxide and TPD mea- resistivity measurements were performed using a differential surements of multilayer films, respectively. four-point probe technique.30,31 The electrodes were formed by deposition of a Au film with a Cr buffer layer using a high- Publication Date (Web): October 2, 2009 | doi: 10.1021/jp904396j 2. Experimental Section vacuum, e-beam, evaporation system and a photolithographic technique. The Si substrates were mounted on a sample holder The graphene oxide films were deposited on SiO2/Si(100) Downloaded by UNIV OF TEXAS AUSTIN on October 2, 2009 | http://pubs.acs.org substrates from an aqueous solution of graphene oxide platelets. stage and heating of the single-layer platelets was performed The nominal thickness of the SiO overlayers was 300 nm. The inside the SEM chamber (Nova NanoSEM600, FEI Co.), which 2 × -6 graphene oxide solutions were formed by exfoliation of graphite has a base pressure of 1 10 Torr. Temperature measure- oxide, which was produced by chemical oxidation of graphite ments were performed with a thermistor that sits below the using the modified Hummers method.13 Two different methods sample (Figure 1a). To correct for the temperature differential are used by our group to exfoliate the platelets: sonication and between the front surface of the Si substrate and the bottom of slow-stirring. The exfoliation is performed on samples with a the sample holder where the thermistor is located, a calibration concentration of 1 mg of graphite oxide per 1 mL of deionized curve was measured with a thermocouple attached to the front water. The sonicated platelets are exfoliated using a 150 W of a blank substrate. Further details of the resistivity measure- 4,11 ultrasonic bath, where the signature for complete exfoliation ment procedure have been described in previous publications. into single-layer graphene oxide platelets is a conversion of the A confocal microscope image of a single-layer graphene oxide suspension from opaque to clear, which typically takes about platelet with patterned electrodes is shown in Figure 3a. 30 min. The platelets produced by this method have an average For the TPD measurements, depositing several droplets on lateral size of ∼0.5 µm.29 Since it is difficult to grow patterned the substrate and letting the suspension dry in air resulted in electrodes onto platelets this small, the method of slow-stirring, multilayer films (Figure 2b). The average thickness of the films which produces platelets with a much larger lateral size, was grown from slow-stirred platelets was measured to be 8.5 ( used for the measurements presented in this study. For the 0.6 nm using ellipsometry (MV-2000, JA Woollam). AFM exfoliation of the slow-stirred platelets, the graphite oxide/water measurements of graphene oxide films grown using this mixture is stirred with a Teflon-coated stir bar for about a week, technique reveal an individual step height of 1 nm for each which only results in partial exfoliation of the platelets. platelet; therefore, the average thickness of the graphene oxide Therefore, the suspension is sedimented, and the clear liquid films grown from slow-stirred platelets is ∼9 monolayers (ML). Reduction Kinetics of Graphene Oxide J. Phys. Chem. C, Vol. xxx, No. xx, XXXX C

Figure 4. (a) Electrical resistivity of an individual graphene oxide sheet reduced at a fixed temperature of 200 °C (inset: scaling of electrical resistivity to molar concentration of oxygen fraction, where graphene oxide before and after reduction is marked with filled circles (b) and amorphous -5 35 j carbon (FC ) 1.6 × 10 Ω · m ) is marked with an open circle (O)) and (b) the relation between mean molar concentration ln C and mean reduction velocity ln Vj, with line of slope n ) 2.2 fit to the high concentration region of plot.

The TPD measurements were performed using a 200 amu, °C. The average value of the resistivity of five different samples quadrupole, residual gas analyzer (RGA) with electron multiplier after thermal treatment at this temperature was 2 ( 0.7 × 10-2 (Hiden HAL 201) mounted in a UHV chamber with a base Ω · m, which is much higher than the value for graphite (1 × pressure of 5 × 10-11 Torr. The ionizer section of the RGA 10-5 Ω · m).32 There are two primary reasons for this. Previous was fitted with a shroud and a collimator that have 6 mm holes XPS measurements of the thermal reduction of graphene oxide in their centers to allow detection of gas molecules evolving provide evidence that a temperature much higher than 200 °C from the surface of the graphene oxide film while preventing is needed to completely remove all or almost all oxygen from the detection of molecules desorbing from other areas of the the surface.26 On the basis of elemental analysis by XPS of sample holder assembly. The samples (13 mm × 13 mm) were similarly prepared thin multilayer films of graphene oxide, the mounted on a molybdenum plate with a hole (9.5 mm diameter) C/O ratio of “as prepared” graphene oxide is 3:1 and is reduced in the center to allow radiative heating of the sample from a down to 9:1 by heating to 500 °C in UHV.26 In addition, the tungsten filament mounted behind the sample holder. Linear TPD measurements detect a loss of carbon during the decom- heating rates were generated using a computer controlled position process (CO2 and CO). Therefore, the reduced graphene proportional feedback control system. Since our TPD measure- oxide is expected to be highly defective, resulting in additional ments indicate that graphene oxide begins to decompose at scattering of charge carriers within the film. temperatures as low as 70 °C, a load lock was used for insertion Since there is a strong correlation between the electrical of samples into the UHV chamber, which permitted the transfer resistivity and stoichiometry of the material, the resistivity can of samples into the chamber without bake out, which would be used to determine the concentration of oxygen in the material. cause partial decomposition of the oxide. To allow temperature For an intrinsic semiconductor, the temperature dependence of measurements of the graphene oxide films, a custom sliding the carrier concentration is given by thermocouple mount was used. The sliding contact consists of - a chromel alumel thermocouple (0.13 mm diameter leads) spot- E welded to the apex of a bent tungsten wire (0.25 mm diameter), η ) √N N exp - g (1) Publication Date (Web): October 2, 2009 | doi: 10.1021/jp904396j i c v ( ) which makes direct contact with the front of the sample. The 2kBT sample holder assembly with sliding thermocouple mount is

Downloaded by UNIV OF TEXAS AUSTIN on October 2, 2009 | http://pubs.acs.org shown in Figure 3b. The primary drawback of this measuring where E is the band gap, k is the Boltzmann constant, and N technique is that the temperature measured by the thermocouple g B c and Nv are the effective density of states at the bottom of the at the point of contact will be less than that of the rest of the conduction band and the top of the valence band, respectively. graphene oxide film because of heat transfer along the thermo- Since graphene oxide is a nonstoichiometric semiconducting couple leads and tungsten spring. To estimate the amount of compound, it is reasonable to assume that over a finite range temperature offset, a tantalum plate was mounted on the sample of concentrations its band gap should depend linearly on the holder and heated until it just began to glow, resulting in a E ∝ C ° molar oxygen concentration ( g ). In fact, recent theoretical measured thermocouple temperature of 475 C. By heating the studies of the dependence of the electrical properties of graphene plate further until the region near the contact point of the oxide on the oxygen to carbon ratio predict a linear increase in tungsten wire began to glow, a temperature of 600 °C was ° band gap as the oxygen concentration increases for graphene measured, which gives a temperature offset of 125 at 475 C. oxide modeled as being composed of epoxide and hydroxyl A linear offset was applied to all of the temperature measure- groups.33,34 Since the resistivity is inversely proportional to the ments to compensate for this effect. Although the repeatability carrier concentration, the change in molar oxygen concentration of the TPD profiles for similar heating rates was found to be ° should depend on the difference in natural logarithms of the within 2 or 3 C of each other, the absolute temperature at TM resistivities is estimated to be accurate to only (20 °C. - ) F- F 3. Results C Cf p(ln ln f) (2) 3.1. Temperature Dependent Resistivity Measurements. The highest temperature used for the reduction of the single- where p is a proportionality factor. The time dependence of the layer graphene oxide in high vacuum (∼10-6 Torr) was 200 resistivity at 200 °C of a single-layer graphene oxide film is shown D J. Phys. Chem. C, Vol. xxx, No. xx, XXXX Jung et al.

Figure 5. (a) Electrical resistivity and the temperature proﬁle of an individual graphene oxide sheet during reduction usinga5Kstepwise change in temperature and (b) an Arrhenius plot from the rate constants, where the slope of the plot revealed an activation energy of 37 ( 1 kcal/mol.

in Figure 4a, where the final resistivity, Ff, is found to be 1 × determination of the activation energy that should be considered 10-2 Ω · m and the initial value of the resistivity before heating is the assumption that the band gap of graphene oxide depends was measured to be 1 × 103 Ω · m. In the inset of this figure, the linearly on the oxygen content. Although the deviation from a molar concentrations of oxygen for graphene oxide are plotted linear behavior has not been directly measured, the good fit of versus the logarithm of the resistivity (filled circles). The molar the data to a linear curve (Figure 5b) provides indirect evidence concentrations are calculated on the basis of the assumption that that this is a valid assumption over the range of reductions the atomic carbon-to-carbon distance in the hexagonal structure measured in this study. of graphene oxide is a ) 0.142 nm and the thickness of graphene 3.2. Temperature Programmed Desorption. Attempts were oxide is d ) 1 nm. The number of atoms in the hexagonal unit made to measure TPD spectra of dispersed single-layer platelets cell is divided by the volume (V ) 33/2a2d), which results in a prepared in a similar way to the samples used in the temperature- concentration at the beginning stage of 2.3 × 104 mol/m3 and a dependent resistivity measurements. However, the partial pres- 3 3 26 value at the final stage of 7.1 × 10 mol/m . sures of H2O, CO2, and CO from the desorbing molecules were In reaction kinetics, the reaction velocity (V) is related to the generally lower than the partial pressures of the corresponding nth power of the concentration of reactant (V)k[C]n), where k residual gases in the UHV chamber. To increase the signal-to- is the rate constant and n is termed the order of the reaction. noise level of the TPD experiment, graphene oxide films The reaction velocity is the rate of change of the concentration, composed of multiple layers of single-layer-thick platelets were V)-dC/dt, for the nonstoichiometric compound. By plotting used. The TPD spectra taken at a heating rate of ) 30 °C/ ln υ versus ln C, the reaction order can be determined from the min for a mass to charge ratio of m/z ) 18, 28, 32, and 44, slope of the curve. As shown in Figure 4b for a single-layer which correspond to the detection of H2O, CO, O2, and CO2, graphene oxide sample annealed at 200 °C, the reaction order are shown in Figure 6a for an ∼9 ML film. To compensate for is estimated to be 2.2 over most of the range of oxygen the difference in ionization efficiency of each molecular species concentrations. Once the reaction order is estimated, the time by the RGA, the measured partial pressures have been divided dependence of the rate constant can be determined using the by the corresponding relative sensitivity of each molecular 36 oxygen concentrations estimated from eq 2. species. Since CO is a cracking fragment of CO2 with a relative 36

Publication Date (Web): October 2, 2009 | doi: 10.1021/jp904396j The temperature dependence of the rate constant for graphene percentage of 8%, this component has been subtracted from oxide is assumed to follow an Arrhenius dependence the measured intensity of the m/z ) 28 partial pressure. To determine the relative amount of each desorbing species from Downloaded by UNIV OF TEXAS AUSTIN on October 2, 2009 | http://pubs.acs.org E films produced from the slow-stirred graphene oxide platelets, k ) A exp(- a ) (3) the areas under the partial pressure versus time curves were RT calculated and are presented in Table 1. The data for the films of slow-stirred platelets were averaged over five samples. As where A is the pre-exponential factor, Ea is the Arrhenius can be seen from the data presented in Figure 6a and Table 1, activation energy, and R is the gas constant. Since the exact the primary desorption component of the samples is H2O, pathway for decomposition is unknown, the Arrhenius activation followed by CO2 and CO. The large H2O component is not energy is interpreted as the energy barrier for the process of unexpected since the films were deposited from aqueous decomposition of the oxide and followed by formation of suspensions, and interlamellar H2O is known to be an integral the desorption products. To obtain the effect of temperature on part of the structure of graphite oxide as well as ‘graphene oxide 37 the rate constant, the resistivity of a single-layer graphene oxide paper’. A comparison of the CO2/CO area ratio for slow-stirred film was monitored usinga5Kstepwise increase in the samples provides a value of 1.9. temperature as shown in Figure 5a. The temperature range was To determine if isolated OH groups or O atoms are desorbing carefully chosen for a relatively slow reduction at the beginning during the decomposition of the graphene oxide, TPD spectra temperature so that a wide range of temperatures could be for m/z ) 17 and 16 were taken. Since the ionization of H2Oin utilized before saturation occurs. To determine Ea, the natural the RGA also produces some OH (21%) and O (2%) and the logarithm of the rate constant is plotted versus inverse temper- ionization of CO2 and CO also produces some O (9% and 2%, ature as shown in Figure 5b. The slope of the curve is the respectively), the measured spectra for m/z ) 17 and 16 must activation energy divided by the gas constant, which results in be corrected for these cracking components. TPD spectra for an activation energy of 37 ( 1 kcal/mol (1.6 eV/molecule) for m/z ) 17, 16, and 15 after the cracking components from H2O, the single-layer graphene oxide platelets. One aspect of the CO2, and CO were subtracted are shown in Figure 6b. The m/z Reduction Kinetics of Graphene Oxide J. Phys. Chem. C, Vol. xxx, No. xx, XXXX E

Figure 6. TPD spectra taken at a heating rate of ) 30 °C/min for ∼9 ML thick ﬁlm of “slow-stirred” graphene oxide (a) for m/z ) 18, 28, 32 and 44, where the inset shows the measured heating rate of the sample, and (b) for m/z ) 17, 16, and 15.

TABLE 1: Areas under Partial Pressure Versus Time Curves for Films of Graphene Oxide

m/zAslow-stirred (Torr · s) 18 2.3 ( 0.4 × 10-6 28 5.3 ( 0.9 × 10-7 32 1.4 ( 0.4 × 10-8 44 1.0 ( 0.1 × 10-6

) 17 spectrum shows only a small broad peak centered at 130 °C, which is most likely the result of a slight deviation of the cracking fraction of OH from H2O for our RGA from the standard value of 21%. However, there is a peak for m/z ) 16 that is centered at 130 °C, which is probably due to methane (CH4) desorption since there is also a small peak for m/z ) 15 2 at this same temperature, which would correspond to the Figure 7. Plot of ln(TM/) versus 1/TM for CO2 desorption measured cracking component CH3. with heating rates of 12, 30, and 59 °C/min for ∼9 ML thick films of The activation energies of the films composed of slow-stirred “slow-stirred” graphene oxide, where the slope of the plot revealed an platelets of graphene oxide can be measured with TPD if one activation energy of 32 ( 4 kcal/mol. assumes an Arrhenius dependence for the probability of desorption. Since the partial pressure of each decomposition therefore, the value of TM will depend on the coverage θ for a fixed product is proportional to the rate of desorption Rdes )-dθ/dt of that species from the surface of the film, where θ is the heating rate . Since the multilayer graphene oxide films in this study are all prepared from the same graphene oxide suspension of slow-

Publication Date (Web): October 2, 2009 | doi: 10.1021/jp904396j instantaneous coverage, the desorption rate can be written as stirred source material, the relative coverage of oxygen and hydrogen in the platelets of the ﬁlms should be the same for each sample. A -E 2 dθ n a plot of ln (TM/) versus 1/TM for CO2 desorption measured with heating Downloaded by UNIV OF TEXAS AUSTIN on October 2, 2009 | http://pubs.acs.org - ) νθ exp( ) (4) dt RT rates of 12, 30, and 59 °C/min is shown in Figure 7. The activation energy found from the slope of this curve is 32 ( 4 kcal/mol (1.4 where ν is the frequency factor.38 If the sample is heated at a constant eV/molecule). rate ) dT/dt, the rate of change of coverage can be written as 4. Discussion n - dθ νθ Ea Since it is relatively easy to exfoliate single-layer graphene )- exp( ) (5) dT RT oxide from graphite oxide, this could be used as a source for nanometer-thick dielectric layers for use in both nanoelectronics and chemical sensor applications. However, the TPD results By taking the derivative with respect to temperature of eq 5 and setting indicate that the decomposition of graphene oxide begins at the it equal to zero, the temperature, TM, at which the maximum of the relatively low temperature of 70 °C, which limits applications partial pressure occurs for a sample heated at constant rate can be 39 on the basis of graphene oxide to approximately room temper- found. This results in the expression ature and below. Another application of graphene oxide that has sparked great interest is the possible use of graphene oxide 2 n-1 as a precursor for the formation of single-layer graphene. Ideally, TM Ea 1 nνRθ ln( ) ) - ln( ) (6) the reduction of graphene oxide would be accomplished by the R TM Ea desorption of the O and OH groups via the formation of O2 and H2O, which would leave the carbon backbone of the The results of the resistivity measurements of single-layer graphene graphene intact. However, the TPD measurements show almost oxide indicate that the order of the decomposition kinetics is ∼2; no O2 formation and a relatively large amount of CO2 and CO F J. Phys. Chem. C, Vol. xxx, No. xx, XXXX Jung et al.

desorbing during the decomposition of the oxide. Since carbon of graphene-based nanosheets via chemical reduction of exfoliated graphite is lost during the reduction process in UHV, the addition of oxide. Carbon 2007, 45, 1558–1565. (7) Paci, J. T.; Belytschko, T.; Schatz, G. C. Computational studies of other carbon-containing reactants during the thermal reduction the structure, behavior upon heating, and mechanical properties of graphite will be needed to regain the electrical and mechanical properties oxide. J. Phys. Chem. C 2007, 111, 18099–18111. of pristine graphene. (8) Kudin, K. N.; Ozbas, B.; Schniepp, H. C.; Prud’homme, R. K.; Aksay, I. A.; Car, R. Raman spectra of graphite oxide and functionalized For TPD measurements of gas molecules adsorbed on single- graphene sheets. Nano Lett. 2008, 8 (1), 36–41. crystal surfaces, multiple desorption peaks are often observed, (9) Becerril, H. A.; Mao, J.; Liu, Z.; Stoltenberg, R.; Bao, Z.; Chen, which results from different binding sites on the surface (on- Y. Evaluation of solution-processed reduced graphene oxide films as top, bridge, 3-fold hollow, etc.). For the TPD measurements of transparent conductors. ACS Nano 2008, 2 (3), 463–470. (10) Eda, G.; Fanchini, G.; Chhowalla, M. Large-area Ultrathin films graphene oxide, the desorption peaks of H2O, CO2, and CO are of reduced graphene oxide as transparent and flexible electronic material. all observed at the same temperature. Since the reduction of Nat. Nanotechnol. 2008, 3, 270–274. graphene oxide is a decomposition process where some of the (11) Jung, I.; Dikin, D. A.; Piner, R. D.; Ruoff, R. S. Tunable electrical - conductivity of individual graphene oxide sheets reduced at “low” temper- carbon backbone is lost, the breaking of carbon carbon bonds atures. Nano Lett. 2008, 8 (12), 4283–4287. must trigger a change in the local chemical potential that is (12) Brodie, B. C. Sur le poids atomique du graphite. Ann. Chim. Phys. strong enough to initiate the simultaneous formation of all of 1860, 59, 466–472. the desorption products. A first-order desorption process is (13) Hummers, W. S., Jr.; Offeman, R. E. Preparation of graphitic oxide. J. Am. Chem. Soc. 1958, 80, 1339. associated with the direct desorption of a particular molecular (14) Staudenmaier, L. Verfahren zur Darstellung der Graphitsäure. Ber. species, whereas a second-order desorption processes involves Dtsch. Chem. Ges. 1898, 31, 1481–1499. an intermediate step before desorption. Therefore, the measured (15) Stankovich, S.; Piner, R. D.; Chen, X.; Wu, N.; Nguyen, S. T.; n ) Ruoff, R. S. Stable aqueous dispersions of graphitic nanoplatelets via value of 2.2 from the temperature-dependent resistivity reduction of exfoliated graphite oxide in the presence of poly(sodium measurements is in general agreement with the expected n ) 2 4-styrenesulfonate). J. Mater. Chem. 2006, 16 (2), 155–156. desorption kinetics for the decomposition of graphene oxide. (16) Stankovich, S.; Piner, R. D.; Nguyen, S. T.; Ruoff, R. S. Synthesis and exfoliation of isocyanate-treated graphene oxide nanoplatelets. Carbon 2006, 44 (15), 3342–3347. 5. Conclusion (17) Szabo, T.; Berkesi, O.; Forgo, P.; Josepovits, K.; Sanakis, Y.; Petridis, D.; De´kańy, I. Evolution of surface functional groups in a series The thermal reduction kinetics of graphene oxide have been of progressively oxidized graphite oxides. Chem. Mater. 2006, 18, 2740– monitored using temperature-dependent resistivity measurements 2749. (18) Hofmann, U.; Holst, R. U¨ ber die Saürenatur und die Methylierung under high vacuum and using TPD in UHV. The exfoliation of von Graphitoxyd. Ber. Dtsch. Chem. Ges. 1939, 72, 754–771. single-layer graphene oxide from graphite oxide has been (19) Lerf, A.; He, H.; Forster, M.; Klinowski, J. Structure of graphite accomplished by slow-stirring, which produces platelets with oxide revisited. J. Phys. Chem. B 1998, 102 (23), 4477–4482. ∼ (20) Ruess, G. U¨ ber das Graphitoxyhydroxyd (Graphitoxyd). Monatsh. an average lateral size of 10 µm. The TPD measurements for Chem. 1946, 76, 381–417. temperatures up to 300 °C find that the primary desorption (21) Scholz, W.; Boehm, H. P. Untersuchungen am Graphitoxyd. Z. components of graphene oxide are H2O, CO2, and CO, with Anorg. Allg. Chem. 1969, 369, 327–340. only trace amounts of O being detected. The detection of CO (22) Nakajima, T.; Matsuo, Y. Formation process and structure of 2 2 graphite oxide. Carbon 1988, 32 (3), 469–475. and CO provides evidence that carbon is being lost from the (23) Cai, W.; Piner, R. D.; Stadermann, F. J.; Park, S.; Shaibat, M. A.; graphene backbone of the oxide during the reduction process. Ishii, Y.; Yang, D.; Velamakanni, A.; An, S. J.; Stoller, M.; An, J.; Chen, The desorption kinetics have been estimated from the temper- D.; Ruoff, R. S. Synthesis and solid-state NMR structural characterization ature-dependent resistivity measurements to be 2.2. For the of 13C-labeled graphite oxide. Science 2008, 321, 1815–1817. (24) Scholtz, W.; Boehm, H.-P. Die thermische Zersetzung von Graphi- decomposition of the slow-stirred graphene oxide, activation toxyd. Naturwissenschaften 1963, 51 (7), 160. energies of 37 ( 1 and 32 ( 4 kcal/mol are measured by (25) Schniepp, H. C.; Li, J. L.; McAllister, M. J.; Sai, H.; Herra-Alonso, resistivity measurements on single-layer graphene oxide and M.; Adamson, D. H.; Prud’homme, R. K.; Car, R.; Saville, D. A.; Aksay, A. Functionalized single graphene sheets derived from spliting graphite Publication Date (Web): October 2, 2009 | doi: 10.1021/jp904396j TPD measurements of multilayer films, respectively. oxide. Phys. Chem. B 2006, 110, 8535–8539. (26) Yang, D.; Velamakanni, A.; Bozoklub, G.; Park, S.; Stoller, M.; Acknowledgment. We gratefully acknowledge support from Piner, R. D.; Stankovich, S.; Jung, I.; Field, D. A.; Ventrice, C. A., Jr.; Downloaded by UNIV OF TEXAS AUSTIN on October 2, 2009 | http://pubs.acs.org the DARPA CERA Center and the Center on Nanoscale Science Ruoff, R. S. Chemical analysis of graphene oxide films after heat and chemical treatments by X-ray photoelectron and micro-raman spectroscopy. and Technology for Integrated Micro/Nano-Electromechanical Carbon 2009, 47, 145–152. Transducers (iMINT) (Award No. HR0011-06-1-0048). (27) Mattevi, C.; Eda, G.; Agnoli, S.; Miller, S.; Mkhoyan, K. A.; Celik, O.; Mastrogiovanni, D.; Granozzi, G.; Garfunkel, E.; Chhowalla, M. Evolution of electrical, chemical, and structural properties of transparent References and Notes and conducting chemically derived graphene thin films. AdV. Funct. Mater. 2009, 19, 2577–2583. (1) Gilje, S.; Han, S.; Wang, M.; Wang, K. L.; Kaner, R. B. A Chemical (28) Boehm, H.-P. Surface chemical characterization of carbons from route to graphene for device applications. Nano Lett. 2007, 7 (11), 3394– adsorption studies. In Adsorption by carbons; Elsevier: Amsterdam, 2008; 3398. pp 301-27. (2) Gomez-Navarro, C.; Thomas Weitz, R. T.; Bittner, A. M.; Scolari, (29) Fan, F. R.; Park, S.; Zhu, Y.; Ruoff, R. S.; Bard, A. J. Electro- M.; Mews, A.; Burghard, M.; Kern, K. Electronic transport properties of generated chemiluminescence of partially oxidized highly oriented pyrolytic individual chemically reduced graphene oxide sheets. Nano Lett. 2007, 7 graphite surfaces and of graphene oxide nanoparticles. J. Am. Chem. Soc. (11), 3499–34503. 2008, 131, 937–939. (3) Si, Y.; Samulski, E. T. Synthesis of water soluble graphen. Nano (30) Van der Pauw, L. J. A method of measuring specific resistivity Lett. 2008, 8 (6), 1679–1682. and Hall effect of discs of arbitrary shape. Philips Res. Reports 1958, 13, (4) Jung, I.; Dikin, D. A.; Park, S.; Cai, W.; Mielke, S. L.; Ruoff, R. S. 1–9. Effect of water vapor on electrical properties of individual reduced graphene (31) Van der Pauw, L. J. A method of measuring the resistivity and oxide sheets. J. Phys. Chem. C 2003, 112 (51), 20264–20268. Hall coefficient on lamellae of arbitrary shape. Philips Tech. ReV. 1958, (5) McAllister, M. J.; Li, J. L.; Adamson, D. H.; Schniepp, H. C.; 20, 220–224. Abdala, A. A.; Liu, J.; Herrera-Alonso, M.; Milius, D. L.; Car, R.; (32) Powell, R. L.; Childs, G. E. American institute of physics handbook, Prud’homme, R. K.; Aksay, I. A. Single sheet functionalized graphene by 3rd ed.; McGraw-Hill: Dallas, 1972; pp 142-160. oxidation and thermal expansion of graphite. Chem. Mater. 2007, 19, 4396– (33) Boukhvalov, D. W.; Katsnelson, M. I. Modeling of graphite oxide. 4404. J. A. Chem. Soc. 2008, 130, 10697–10701. (6) Stankovich, S.; Dikin, D. A.; Piner, R. D.; Kohlhaas, K. A.; (34) Yan, J.-A.; Xian, L.; Chou, M. Y. Structural and electronic Kleinhammes, A.; Jia, Y.; Wu, Y.; Nguyen, S. T.; Ruoff, R. S. Synthesis properties of oxidized graphene. Phys. ReV. Lett. 2009, 103, 086802. Reduction Kinetics of Graphene Oxide J. Phys. Chem. C, Vol. xxx, No. xx, XXXX G

(35) Stransﬁeld, A. The electric furnace: its construction, operation and (38) Polanyi, M.; Wigner, E. Bildung und Zerfall von Moleku¨len. Z. uses; McGraw-Hill: New York, 1914; p 89. Phys. 1925, 33 (1), 429–434. (36) Hiden Analytical (http://www.hidenanalytical.com/index.php/en/ (39) Redhead, P. A. Thermal desorption of gases. Vacuum 1962, 12, cracking-patterns). 203–211. (37) Dikin, D. A.; Stankovich, S.; Zimney, E. J.; Piner, R. D.; Dommett, H. B.; Evmenenko, G.; Nguyen, S. T.; Ruoff, S. R. Preparation and characterization of graphene oxide paper. Nature 2007, 448, 457–460. JP904396J Publication Date (Web): October 2, 2009 | doi: 10.1021/jp904396j Downloaded by UNIV OF TEXAS AUSTIN on October 2, 2009 | http://pubs.acs.org